Plant lipids have a variety of industrial and nutritional uses and are central to plant membrane function and climatic adaptation. These lipids represent a vast array of chemical structures, and these structures determine the physiological and industrial properties of the lipid. Many of these structures result either directly or indirectly from metabolic processes that alter the degree of unsaturation of the lipid. Different metabolic regimes in different plants produce these altered lipids, and either domestication of exotic plant species or modification of agronomically adapted species is usually required to economically produce large amounts of the desired lipid.
Plant lipids find their major use as edible oils in the form of triacylglycerols. The specific performance and health attributes of edible oils are determined largely by their fatty acid composition. Most vegetable oils derived from commercial plant varieties are composed primarily of palmitic (16:0), stearic (18:0), oleic (18:1), linoleic (18:2) and linolenic (18:3) acids. Palmitic and stearic acids are, respectively, 16- and 18-carbon-long, saturated fatty acids. Oleic, linoleic, and linolenic acids are 18-carbon-long, unsaturated fatty acids containing one, two, and three double bonds, respectively. Oleic acid is referred to as a mono-unsaturated fatty acid, while linoleic and linolenic acids are referred to as poly-unsaturated fatty acids. The relative amounts of saturated and unsaturated fatty acids in commonly used, edible vegetable oils are summarized below (Table 1):
TABLE 1Percentages of Saturated and UnsaturatedFatty Acids in the Oils of Selected Oil CropsMono-Poly-SaturatedunsaturatedunsaturatedCanola 6%58%36%Soybean15%24%61%Corn13%25%62%Peanut18%48%34%Safflower 9%13%78%Sunflower 9%41%51%Cotton30%19%51%
Many recent research efforts have examined the role that saturated and unsaturated fatty acids play in reducing the risk of coronary heart disease. In the past, it was believed that mono-unsaturates, in contrast to saturates and poly-unsaturates, had no effect on serum cholesterol and coronary heart disease risk. Several recent human clinical studies suggest that diets high in mono-unsaturated fat and low in saturated fat may reduce the “bad” (low-density lipoprotein) cholesterol while maintaining the “good” (high-density lipoprotein) cholesterol (Mattson et al., Journal of Lipid Research (1985) 26:194–202).
A vegetable oil low in total saturates and high in mono-unsaturates would provide significant health benefits to consumers as well as economic benefits to oil processors. As an example, canola oil is considered a very healthy oil. However, in use, the high level of poly-unsaturated fatty acids in canola oil renders the oil unstable, easily oxidized, and susceptible to development of disagreeable odors and flavors (Gailliard, 1980, Vol. 4, pp. 85–116 In: Stumpf, P. K., Ed., The Biochemistry of Plants, Academic Press, New York). The levels of poly-unsaturates may be reduced by hydrogenation, but the expense of this process and the concomitant production of nutritionally questionable trans isomers of the remaining unsaturated fatty acids reduces the overall desirability of the hydrogenated oil (Mensink et al., New England J. Medicine (1990) N323: 439–445). Similar problems exist with soybean and corn oils.
For specialized uses, high levels of poly-unsaturates can be desirable. Linoleate and linolenate are essential fatty acids in human diets, and an edible oil high in these fatty acids can be used for nutritional supplements, for example in baby foods.
Mutation-breeding programs have met with some success in altering the levels of poly-unsaturated fatty acid levels found in the edible oils of agronomic species. Examples of commercially grown varieties are high (85%) oleic sunflower and low (2%) linolenic flax (Knowles, (1980) pp. 35–38 In: Applewhite, T. H., Ed., World Conference on Biotechnology for the Fats and Oils Industry Proceedings, American Oil Chemists' Society). Similar commercial progress with the other plants shown in Table 1 has been largely elusive due to the difficult nature of the procedure and the pleiotropic effects of the mutational regime on plant hardiness and yield potential.
The biosynthesis of the major plant lipids has been the focus of much research (Browse et al., Ann. Rev. Plant Physiol. Mol. Biol. (1991) 42:467–506). These studies show that, with the notable exception of the soluble stearoyl-acyl carrier protein desaturase, the controlling steps in the production of unsaturated fatty acids are largely catalyzed by membrane-associated fatty acid desaturases. Desaturation reactions occur in plastids and in the endoplasmic reticulum using a variety of substrates including galactolipids, sulfo-lipids, and phospholipids. Genetic and-physiological analyses of Arabidopsis thaliana nuclear mutants defective in various fatty acid desaturation reactions indicates that most of these reactions are catalyzed by enzymes encoded at single genetic loci in the plant. The analyses show further that the different defects in fatty acid desaturation can have profound and different effects on the ultra-structural morphology, cold sensitivity, and photosynthetic capacity of the plants (Ohlrogge, et al., Biochim. Biophys. Acta (1991) 1082:1–26). However, biochemical characterization of the desaturase reactions has been meager. The instability of the enzymes and the intractability of their proper assay has largely limited researchers to investigations of enzyme activities in crude membrane preparations. These investigations have, however, demonstrated the role of delta-12 desaturase and delta-15 desaturase activities in the production of linoleate and linolenate from 2-oleoyl-phosphatidyl-choline and 2-linoleoyl-phosphatidylcholine, respectively (Wang et al., Plant Physiol. Biochem. (1988) 26:777–792). Thus, modification of the activities of these enzymes represents an attractive target for altering the levels of lipid unsaturation by genetic engineering.
Nucleotide sequences encoding microsomal delta-9 stearoyl-coenzyme-A desaturases from yeast, rat, and mice have been described (Stukey, et al., J. Biol. Chem.(1990) 265:20144–20149; Thiede, et al., J. Biol. Chem. (1986) 261:13230–13235; Kaestner, et al., J. Biol. Chem. (1989) 264:14755–1476). Nucleotide sequences encoding soluble delta-9 stearoyl-acyl carrier protein desaturases from higher plants have also been described (Thompson, et al., Proc. Natl. Acad. Sci. U.S.A. (1991) 88:2578–2582; Shanklin et al., Proc. Natl. Acad. Sci. USA (1991) 88:2510–2514). A nucleotide sequence from coriander plant encoding a soluble fatty acid desaturase, whose deduced amino acid sequence is highly identical to that of the stearoyl-acyl carrier protein desaturase and which is responsible for introducing the double bond in petroselinic fatty acid (18:1, 6c), has also been described (Cahoon, et. al. (1992) Proc. Natl. Acad. Sci. U.S.A. 89:11184–11188). Two fatty acid desaturase genes from the cyanobacterium, Synechocystis PCC6803, have been described: one encodes a fatty acid desaturase, designated des A, that catalyzes the conversion of oleic acid at the sn-1 position of galactolipids to linoleic acid [Wada, et al., Nature (1990) 347:200–203]; another encodes a delta-6 fatty acid desaturase that catalyzes the conversion of linoleic acid at the sn-1 position of galactolipids to γ-linolenic acid (18:2, 6c,9c) [WO 9306712]. Nucleotide sequences encoding higher plant membrane-bound microsomal and plastid delta-15 fatty acid desaturases have also been described [WO 9311245]; Arondel, V. et. al. (1992) Science 258:1353–1355]. There is no report of the isolation of higher plant genes encoding fatty acid desaturases other than the soluble delta-6 and delta-9 desaturases and the membrane-bound (microsomal and plastid) delta-15 desaturases. While there is extensive amino acid sequence identity between the soluble desaturases and significant amino acid sequence identity between the higher plant microsomal and plastid delta-15 desaturases, there is no significant homology between the soluble and the membrane-bound desaturases. Sequence-dependent protocols based on the sequences encoding delta-15 desaturases have been unsuccessful in cloning sequences for microsomal delta-12 desaturase. For example, nucleotide sequences of microsomal or plastid delta-15 desaturases as hybridization probes have been unsuccessful in isolating a plant microsomal delta-12 desaturase clone. Furthermore, while we have used a set of degenerate oligomers made to a stretch of 12 amino acids, which is identical in all plant delta-15 desaturases and highly conserved (10/12) in the cyanobacterial des A desaturase, as a hybridization probe to isolate a higher plant nucleotide sequence encoding plastid delta-12 fatty acid desaturase, this method has been unsuccessful in isolating the microsomal delta-12 desaturase cDNAs. Furthermore, there has been no success in isolating the microsomal delta-12 desaturase by using the polymerase chain reaction products derived from plant DNA, plant RNA or plant cDNA library using PCR primers made to stretches of amino acids that are conserved between the higher plant delta-15 and des A desaturases. Thus, there are no teachings which enable the isolation of plant microsomal delta-12 fatty acid desaturases or plant fatty acid desaturase-related enzymes. Furthermore, there is no evidence for a method to control the level of delta-12 fatty acid desaturation or hydroxlylation in plants using nucleic acids encoding delta-12 fatty acid desaturases or hydroxylases.
The biosynthesis of the minor plant lipids has been less well studied. While hundreds of different fatty acids have been found, many from the plant kingdom, only a tiny fraction of all plants have been surveyed for their lipid content (Gunstone, et al., Eds., (1986) The Lipids Handbook, Chapman and Hall Ltd., Cambridge). Accordingly, little is known about the biosynthesis of these unusual fatty acids and fatty acid derivatives. Interesting chemical features found in such fatty acids include, for example, allenic and conjugated double bonds, acetylenic bonds, trans double bonds, multiple double bonds, and single double bonds in a wide number of positions and configurations along the fatty acid chain. Similarly, many of the structural modifications found in unusual lipids (e.g., hydroxylation, epoxidation, cyclization, etc.) are probably produced via further metabolism following chemical activation of the fatty acid by desaturation or they involve a chemical reaction that is mechanistically similar to desaturation. Many of these fatty acids and derivatives having such features within their structure could prove commercially useful if an agronomically viable species could be induced to synthesize them by introduction of a gene encoding the appropriate desaturase. Of particular interest are vegetable oils rich in 12-hydroxyoctadeca-9-enoic acid (ricinoleic acid). Ricinoleic acid and its derivatives are widely used in the manufacture of lubricants, polymers, cosmetics, coatings and pharmaceuticals (e.g., see Gunstone, et al., Eds., (1986) The Lipids Handbook, Chapman and Hall Ltd., Cambridge). The only commercial source of ricinoleic acid is castor oil and 100% of the castor oil used by the U.S. is derived from beans grown elsewhere in the world, mainly Brazil. Ricinoleic acid in castor beans is synthesized by the addition of an hydroxyl group at the delta-12 position of oleic acid (Galliard & Stumpf (1966) J. Biol. Chem. 241: 5806–5812). This reaction resembles the initial reaction in a possible mechanism for the desaturation of oleate at the delta-12 position to linoleate since dehydration of 12-hydroxyoctadeca-9-enoic acid, by an enzyme activity analogous to the hydroxydecanoyl dehydrase of E. coli (Cronan et al. (1988) J. Biol. Chem. 263:4641–4646), would result in the formation of linoleic acid. Evidence for the hydroxylation reaction being part of a general mechanism of enzyme-catalyzed desaturation in eukaryotes has been obtained by substituting a sulfur atom in the place of carbon at the delta-9 position of stearic acid. When incubated with yeast cell extracts the thiostearate was converted to a 9-sulfoxide (Buist et al. (1987) Tetrahedron Letters 28:857–860). This sulfoxidation was specific for sulfur at the delta-9 position and did not occur in a yeast delta-9-desaturase deficient mutant (Buist & Marecak (1991) Tetrahedron Letters 32:891–894). The 9-sulfoxide is the sulfur analogue of 9-hydroxyocta-decastearate, the proposed intermediate of stearate desaturation.
Hydroxylation of oleic acid to ricinoleic acid in castor bean cells, like microsomal desaturation of oleate in plants, occurs at the delta-12 position of the fatty acid at the sn-2 position of phosphatidylcholine in microsomes (Bafor et al. (1991) Plant Physiol 280:507–514). Furthermore, castor oleate delta-12 hydroxylation and plant oleate microsomal delta-12 desaturation are both inhibited by iron chelators and require molecular oxygen [Moreau & Stumpf (1981) Plant Physiology 67:672–676; Somerville, C. (1992) MSU-DOE Plant Research Laboratory Annual Report]. These biochemical similarities in conjunction with the observation that antibodies raised against cytochrome b5 completely inhibit the activities of both oleate delta-12 desaturation in safflower microsomes and oleate delta-12 hydroxylase in castor microsomes [Somerville, C. (1992) MSU-DOE Plant Research Laboratory Annual Report] comprise strong evidence that the hydroxylase and the desaturase are functionally related. It seems reasonable to assume, therefore, that the nucleotide sequence encoding a plant delta-12 desaturase would be useful in cloning the oleate hydroxylase gene from castor by sequence-dependent protocols. For example, by screening a castor DNA library with oligomers based on amino acid regions conserved between delta-12 desaturases, or regions conserved between delta-12 and other desaturases, or with oligomers based on amino acids conserved between delta-12 desaturases and known membrane-associated hydroxylases. It would be more efficient to isolate the castor oleate hydroxylase cDNA by combining the sequence dependent protocols with a “differential” library approach. One example of such a difference library would be based on different stages of castor seed development, since ricinoleic acid is not synthesized by very young castor seeds (less than 12 DAP, corresponding to stage I and stage II seeds in the scheme of Greenwood & Bewley, Can. J. Bot. (1982) 60:1751–1760), in the 20 days following these early stages the relative ricinoleate content increases from 0% to almost 90% of total seed fatty acids (James et al. Biochem. J. (1965) 95:448–452, Canvin. Can. J. Biochem. Physiol. (1963) 41:1879–1885). Thus it would be possible to make a cDNA “difference” library made from mRNA present in a stage when ricinoleic acid was being synthesized at a high rate but from which mRNA present in earlier stages was removed. For the earlier stage mRNA, a stage such as stage II (10 DAP) when ricinoleic acid is not being made but when other unsaturated fatty acids are, would be appropriate. The construction of libraries containing only differentially expressed genes is well known in the art (Sargent. Meth. Enzymol. (1987) 152:423–432). Assembly of the free ricinoleic acid, via ricinoleoyl-CoA, into triacylglycerol is readily catalyzed by canola and safflower seed microsomes (Bafor et al., Biochem J. (1991) 280:507–514, Wiberg et al. 10th International Symposium on the Metabolism, Strucure & Function of Plant Lipids (1992), Jerba, Tunisia) and ricinoleic acid is removed from phosphatidylcholine by a lipase common to all oilseeds investigated. Thus, expression of the castor bean oleate hydroxylase gene in oil crops, such as canola seeds and soybeans, would be expected to result in an oil rich in triglycerides containing ricinoleic acid.